INDEX

Explanations

references to software or technology-related features

oai_token-act-pair · gpt-3.5-turbo

mentions of new or notable features in technology or software

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alg

-0.81

amn

-0.74

oner

-0.74

externalActionCode

-0.72

Ã§a

-0.71

sis

-0.70

apy

-0.69

cffffcc

-0.69

thur

-0.67

uin

-0.67

POSITIVE LOGITS

Features

0.93

 features

0.91

 Features

0.88

Feature

0.85

iveness

0.82

afety

0.81

ivity

0.79

 Include

0.75

 enhancements

0.75

 prominently

0.71

Activations Density 0.026%

references to software or technology-related features

mentions of new or notable features in technology or software

No Comments

No Known Activations